Nokia today announced the industry’s first commercial AI-RAN platform, marking one of the most significant changes to radio network architecture in decades. As AI establishes itself as the primary workload in mobile networks, telecom providers need greater capacity, greater profitability, and faster innovation without relying on traditional hardware upgrade cycles.
Nokia’s AI-RAN platform helps telecom providers unlock significantly more uplink and downlink capacity from the spectrum and radio infrastructure they already own, offering a practical path to AI-native networks, while improving network profitability and accelerating innovation at the speed of software.
Based on Nokia’s AI-native network architecture and NVIDIA accelerated computing, Nokia’s AI-RAN platform delivers a quantum leap in network performance and cost-effectiveness. The AI-RAN platform has already demonstrated a greater than 20% increase in spectral efficiency thanks to AI-powered radio innovations. The company plans to achieve a 50% increase in spectral efficiency by 2027 and greater than 100% by 2028, enabling telecom providers to handle significantly higher traffic volumes in dense cells, while reducing cost per bit and improving customer experience.
Through a new software subscription model, telecom providers can benefit from AI innovation, new features and performance improvements without relying on hardware renewal cycles. Nokia’s AI-RAN solutions will enter pilot phase later this year and will be commercially available in 2027, with a roadmap that leverages NVIDIA’s programmable silicon platforms.
A native AI platform. Three ways to adopt
Recognizing the diversity of network strategies and installed RAN bases of telecom providers, Nokia’s AI-RAN platform is based on a common software-defined architecture, powered by Nokia anyRAN software and NVIDIA accelerated computing. Compatible with 4G, 5G and future network evolution, it offers three hardware platform options, including an expansion card for existing AirScale deployments and a Cloud RAN alternative. Fully compatible with Open RAN standards, the platform supports open and interoperable deployments from multiple vendors, giving operators the flexibility to choose the hardware and cloud environments that best suit their needs. These hardware platform options allow telecom providers to modernize at their own pace, preserving existing infrastructure investments, benefiting from a common software roadmap, and accelerating innovation at the speed of software. Telecom providers can adopt AI-RAN in stages, using the approach that best fits their deployment strategy, capacity requirements and installed base.
Leverage existing investments
For current Nokia customers, the company presents the new AirScale expansion unit with GPU as the most efficient solution for the future. Designed for Nokia’s installed AirScale infrastructure, this solution integrates NVIDIA accelerated computing into existing network infrastructure, enabling a significant increase in capacity through a simple upgrade while preserving network investments. This approach is also supported by commercial AI-accelerated chips from Marvell, as part of Nokia’s broader ecosystem for software-defined AI-RAN architectures. Telecom providers can incorporate advanced AI capabilities, continuously improve performance through software, and extend the value of the deployed infrastructure.
Expand native AI capability anywhere
For telecom providers looking for maximum flexibility and performance in deployment, Nokia introduces the industry’s first standalone AI-RAN node with GPU. This node brings AI-accelerated RAN performance to any network environment and supports 4G, 5G, and future 6G workloads on a common platform. The solution can be deployed as a standalone node, in clustered configurations, or alongside AirScale as a single logical base station, providing operators with a highly flexible path to scale AI-native networks without compromising deployment freedom.
Enabling cloud-native AI-RAN
For telecom providers adopting cloud-native architectures, Nokia introduces COTS server solutions for AI-RAN with GPUs, distributed through ecosystem partners. These platforms enable an open and secure supply chain while supporting deployment on industry-standard accelerated computing infrastructures, combining the flexibility of the cloud with the performance requirements of AI radio networks.
Innovation at the speed of software
Nokia’s AI-RAN technology marks a fundamental shift from hardware-defined radio networks to software-defined platforms that continually improve through innovation in software and artificial intelligence. Through Nokia’s new subscription business model, telecom providers gain continuous access to advanced AI algorithms, improvements in spectral efficiency, network optimization capabilities and future native AI functions that can be enabled by software. This approach allows them to benefit from continuous innovation, while maximizing the long-term return on investment in infrastructure, resulting in lower total cost of ownership and better performance without hardware premium. Instead of waiting for the next hardware cycle, networks can continually improve their performance, efficiency, security, and resiliency as new capabilities become available.
Nokia introduces the industry’s first standalone AI-RAN node with GPU. This node offers AI-accelerated RAN performance
By combining AI-accelerated computing, advanced AI algorithms and an open ecosystem approach, Nokia is helping telecom providers unlock greater capacity, stronger economics and continuous innovation, while laying the foundation for future network evolution.
